Translation method based on intelligent glasses

Abstract

The invention discloses a translation method based on intelligent glasses, relates to the technical field of real-time translation of glasses, and is used for solving the problem that the actual use experience is affected due to the fact that translation content is not matched with the attention degree of a user. Judging whether to start a translation analysis system according to the recognition result and the native language of the user; monitoring the residence time of the user and the static duration of eyeballs by the system to generate translation characteristic values; judging whether to execute translation according to the distance variation between the user and the article; collecting target state data within the set translation time; and calculating a shallow time proofreading proportion in combination with a translation language length, and finally generating translation shallow time, thereby realizing intelligent recognition of translation requirements and dynamic adjustment of translation duration.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of real-time translation of glasses, more particularly, to a translation method based on smart glasses. BACKGROUND

[0002] At present, most of the traditional translation devices rely on the user to actively input or shoot the target content to trigger the translation process, and lack the intelligent perception ability of the user's real language demand and interactive intention, especially in the real environment with mixed languages (such as airports, shopping malls, exhibitions, etc.), the user often cannot quickly lock the target information for efficient translation.

[0003] The prior art has the following disadvantages: At present, the existing system generally fails to combine the key behavior parameters such as the mother tongue information of the user, the gaze behavior, and the spatial distance change with the target object, cannot achieve the dynamic judgment of the translation demand and the individualized regulation of the display time length, leading to the mismatch between the translation content and the user's attention level, and affecting the actual use experience. Therefore, a translation method based on smart glasses is proposed.

[0004] The above information disclosed in the background section is only intended to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute the prior art known to those of ordinary skill in the art. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a translation method based on smart glasses, which solves the problems raised in the above background technology by using the joint processing mode of user language information modeling, environmental semantic recognition, gaze behavior detection and translation shallow time dynamic adjustment mechanism.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical scheme, a translation method based on smart glasses, comprising the following steps: Step S1: receiving and recording the user input mother tongue category, marking the objects in the current environment by the recognition accessories in the glasses, identifying the information of the marked objects, and judging whether to enter the translation analysis system according to the information identification result and the recorded mother tongue category; Step S2: when entering the translation analysis system, monitoring the residence time of the current environment and the eye ball static time length, generating the translation characteristic value of the current environment by fusing the residence time and the eye ball static time length, monitoring the distance between the user and the marked object and counting the distance change amount of the marked object; Step S3: judging whether to translate the marked object by comprehensively considering the translation characteristic value of the current environment and the distance change amount of the marked object, setting the translation time, and detecting the state data of the marked object within the translation time; Step S4: obtaining a translation flattening trend according to the state data of the marked item, detecting a translation language length of the translation column, calling a translation flattening time benchmark value, calculating a translation flattening time correction ratio by comprehensively considering the translation language length of the translation column and the translation flattening trend, and generating the translation flattening time by combining the translation flattening time benchmark value.

[0007] In a preferred embodiment, in step S1, the text content information of the marked item is identified, the text content information is matched with language categories, and the language categories of the marked item are obtained. If the language categories of the marked item are inconsistent with the native language categories input by the user, the translation analysis system is entered, otherwise, the translation analysis system is not entered.

[0008] In a preferred embodiment, in step S2, the time point when the translation analysis system is entered is taken as a stagnation start time point. A preset time interval is set, the head stable angle at the start and end time of the time interval is collected, and the angle change rate is taken as the ratio of the difference between the head stable angles at the start and end time of the time interval to the time interval. If the angle change rate exceeds a preset angle change rate threshold, it is determined that the user's stagnation is over, and the stagnation stop time point is taken as the stagnation stop time point. The difference between the stagnation start time point and the stagnation stop time point is the stagnation time of the current environment.

[0009] In a preferred embodiment, in step S2, the movement state of the user's eyeball is monitored in real time, and it is determined whether the user's eyeball is in a stationary gaze state. When the movement of the eyeball is monitored, the time point when the movement occurs is recorded. The difference between the stagnation start time point and the time point when the movement occurs is taken as the eyeball stationary duration. After the stagnation time of the current environment and the eyeball stationary duration are standardized, the standardized stagnation time and the standardized eyeball stationary duration of the current environment are obtained, respectively. The standardized stagnation time and the standardized eyeball stationary duration of the current environment are fused to generate the translation feature value of the current environment through a generalized mean aggregation operator.

[0010] In a preferred embodiment, in step S2, the distance between the user and the marked item is monitored in real time, and the corresponding marking time point is recorded when the identified item is marked, and the distance between the user and the marked item at the marking time point is taken as the initial distance. The distance between the user and the marked item at the current time is taken as the current distance. The difference between the initial distance and the current distance is taken as the distance change of the marked item.

[0011] In a preferred embodiment, in step S3, the translation feature value of the current environment and the distance change amount of the marked article are normalized. After the normalization, the translation feature value of the current environment and the normalized result of the distance change amount of the marked article are obtained as input variables of the logistic regression model, and the translation probability value of the marked article is output by the logistic regression model.

[0012] In a preferred embodiment, in step S3, when the translation probability value of the marked article is greater than or equal to a preset translation determination threshold, it is determined that the marked article is translated. On the contrary, it is determined that the marked article is not translated.

[0013] In a preferred embodiment, in step S3, the moment when it is determined that the marked article is translated is taken as the start moment of the preset time window, and the preset time window is set as the translation time. In the translation time, the pixel area of the text region in the image is extracted, and the ratio of the pixel area of the text region to the total pixel area of the image is taken as the text coverage rate. Based on the image center position coordinate difference of the marked article in the translation time and the translation time, the linear velocity is calculated, and the article moving speed is obtained.

[0014] In a preferred embodiment, in step S4, the article moving speed and the text coverage rate are integrated to obtain the translation shallowing trend. The number of characters in the translation column output text is counted as the translation language length. The translation language length and the translation shallowing trend are taken as input variables, and the translation shallowing time correction ratio is calculated by using the normalized compression division function. The translation shallowing time correction ratio is multiplied by a preset translation shallowing time reference value to obtain the translation shallowing time of the marked article.

[0015] Technical effects and advantages of the present application: 1.The present application marks and identifies information by receiving and recording the native language category of the user input, the object in the current environment, judges whether to enter the translation analysis system according to the information identification result and the recorded native language category, generates the translation characteristic value of the current environment after entering the translation analysis system, monitors the residence time of the current environment and the length of time when the eyeball is stationary, monitors the distance between the user and the marked object and statistics the distance change of the marked object, judges whether to translate the marked object by comprehensively judging the translation characteristic value of the current environment, detects the state data of the marked object within the set translation time, obtains the translation shallow trend, detects the translation language length of the translation column, calculates the translation shallow time correction ratio by comprehensively calculating the translation shallow trend, generates the translation shallow time combined with the translation shallow time benchmark value, realizes the intelligent judgment of translation demand and the dynamic adjustment of translation display time length, further, the present application realizes the translation decision mechanism combined with the semantic perception of the environment and the user interaction behavior characteristics, not only can accurately identify the attention degree of the user to the target object, but also can realize the adaptive shallowization and display rhythm optimization of the translation content according to the length of the translation column language output, the user's gaze time and the language complexity. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 The implementation flowchart of the present application is a translation method based on smart glasses.

[0017] Figure 2 The step schematic diagram of the present application is a translation method based on smart glasses. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0019] The application records the mother tongue of the user, marks the objects in the current environment, and performs information recognition, judges whether to enter the translation analysis system according to the information recognition result and the recorded mother tongue, monitors the residence time of the current environment and the length of time when the eyeball is stationary after entering the translation analysis system, generates the translation characteristic value of the current environment, monitors the distance between the user and the marked object and the distance change of the marked object, judges whether to translate the marked object according to the translation characteristic value of the current environment, detects the state data of the marked object within the set translation time, obtains the translation shallowing trend, detects the translation language length of the translation column, calculates the translation shallowing time correction ratio according to the translation language length of the translation column and the translation shallowing trend, and generates the translation shallowing time combined with the translation shallowing time reference value, so as to realize the intelligent judgment of the translation demand and the dynamic adjustment of the translation display time.

[0020] Embodiment 1, a translation method based on intelligent glasses, as shown in the figure, comprising the following steps: Figure 1 Step S1: receiving and recording the mother tongue input by the user, marking the objects in the current environment by the recognition accessory in the glasses, performing information recognition on the marked objects, and judging whether to enter the translation analysis system according to the information recognition result and the recorded mother tongue; Step S2: when entering the translation analysis system, monitoring the residence time of the current environment and the length of time when the eyeball is stationary, generating the translation characteristic value of the current environment by fusing the residence time and the length of time when the eyeball is stationary, and monitoring the distance between the user and the marked object and the distance change of the marked object; Step S3: judging whether to translate the marked object according to the translation characteristic value of the current environment and the distance change of the marked object, setting the translation time, and detecting the state data of the marked object within the translation time; Step S4: obtaining the translation shallowing trend according to the state data of the marked object, detecting the translation language length of the translation column, calling the translation shallowing time reference value, calculating the translation shallowing time correction ratio according to the translation language length of the translation column and the translation shallowing trend, and generating the translation shallowing time combined with the translation shallowing time reference value. Specific steps are as follows:

[0021] In step S1, the mother tongue input by the user is obtained and recorded through the voice recognition accessory, and the images of the environment where the user is located are collected in real time through the camera accessory, and the objects in the images are identified and marked; The identified text content information of the marked object is matched with the language type through the multilingual recognition tool to obtain the language type of the marked object; If the language type of the marked object is inconsistent with the mother tongue input by the user, the translation analysis system is entered, otherwise, the translation analysis system is not entered; If the language type of the marked object is inconsistent with the mother tongue input by the user, the translation analysis system is entered, otherwise, the translation analysis system is not entered; The voice recognition accessory receives and records the native language category input by the user, the camera accessory collects and marks the image of the environment where the user is currently located, realizes the dynamic comparison between the language background of the user and the objects in the actual environment, and judges whether to start the translation analysis system based on the language difference, thereby avoiding invalid processing of content that does not need to be translated, effectively reducing system resource consumption, improving the intelligence and pertinence of the translation process, and improving the accuracy and real-time of user interaction experience.

[0022] It should be noted that the voice recognition accessory is a sound input and voice recognition processing component integrated in the smart glasses, which converts the collected voice signal into text data and identifies the language category corresponding to the voice; the camera accessory is used to capture the environmental image in the wearer's field of view and transmit it to the processing system, which is used to collect the image of the environment where the user is located in this example, and detect and mark the objects in the image; the multilingual recognition tool is a software module for judging the language category of the text content, which can identify different language categories.

[0023] By fusing the user's language background and environmental language information, accurate determination of translation needs is realized, effectively avoiding invalid translation processing of non-target language content, significantly reducing system calculation and resource consumption. At the same time, dynamically starting the translation analysis system improves the real-time and intelligent level of translation response, so that the smart glasses can more accurately capture the translation object of user's attention, enhance the individualization and pertinence of user interaction experience. This technical scheme not only improves the information acquisition efficiency in a multilingual environment, but also greatly improves the practicability and user satisfaction of the system, and is suitable for various application scenarios such as tourism, business and cross-language communication.

[0024] In step S2, the time point when entering the translation analysis system is taken as the stagnation start time point, a preset time interval, the head stable angle at the start and end time of the time interval is collected by the head inertia sensor, and the ratio of the difference between the head stable angles at the start and end time of the time interval to the time interval is taken as the angle change rate; If the angle change rate exceeds the preset angle change rate threshold, it is judged that the user stagnation is over, and the user stagnation end time point is taken as the stagnation stop time point; The difference between the stagnation start time point and the stagnation stop time point is the stagnation time of the current environment.

[0025] The eye tracking accessory collects the motion state of the user's eyeball in real time, judges whether the user's eyeball is in a stationary gaze state, and records the time point when the eyeball moves when the movement is monitored; The difference between the stagnation start time point and the time point when the movement occurs is taken as the eyeball stationary duration.

[0026] The current environment's retention time and eye ball stationary time length are standardized to obtain the current environment's standardized retention time and standardized eye ball stationary time length respectively; The current environment's standardized retention time and standardized eye ball stationary time length are fused to generate the current environment's translation feature value by a generalized mean aggregation operator: wherein, is a preset adjustment parameter, is the current environment's standardized retention time, is the standardized eye ball stationary time length, is the current environment's translation feature value; It needs to be explained that the preset adjustment parameter is used to control the sensitivity of the fused feature. If the preset adjustment parameter is set to 1, it means that the weight of the current environment's retention time and eye ball stationary time length is equal. If the preset adjustment parameter is set to a larger value, it means that the current environment's translation feature value tends to be a larger value of the input variable. If the preset adjustment parameter is set to a smaller value, it means that the current environment's translation feature value tends to be a smaller value of the input variable. The value of the preset adjustment parameter is set by professional personnel according to the actual application scene and user behavior characteristics.

[0027] The distance between the user and the marked object is monitored in real time by an image depth perception sensor. The corresponding marking time point is recorded when the object is identified and marked. The distance between the user and the marked object at the marking time point is taken as the initial distance, and the distance between the user and the marked object at the current time is taken as the current distance. The difference between the initial distance and the current distance is taken as the distance change of the marked object. When the distance change is positive, it means that the user is approaching the marked object. When the distance change is negative, it means that the user is moving away from the marked object. When the distance change is zero, it means that the user's position is stable.

[0028] By fusing the gaze stability and the dynamic change of spatial behavior, a real-time translation judgment strategy is realized, which effectively improves the response accuracy and user experience of the translation process, and reduces the system false trigger rate and resource waste.

[0029] It should be noted that the preset time interval refers to a fixed length of time window set by the system, which is used to count the attitude change of the user's head in the time period; the head inertia sensor is an inertial measurement unit in the smart glasses, which is used to perceive and measure the rotation angle of the user's head in real time; the eye tracking accessory is a sensing module of the smart glasses, which captures the motion trajectory and gaze point of the user's eyes, and in this case, is used to determine whether the eyes are in a stationary gaze state; the preset angle change rate threshold is used to determine the angle range of the stable state of the user's head, which is set by professionals based on the natural shaking range of the human head and application requirements, to ensure accurate distinction between stable and moving states; the image depth perception sensor is a sensing module that can capture the depth information of the three-dimensional space of the environment, which assists in realizing real-time measurement of the spatial distance between the marked object and the user by analyzing the distance information of each point in the scene; the generalized mean aggregation operator is a function for integrating multiple input values into a single characteristic value, and in the present application, the generalized mean aggregation operator is used to fuse the user's residence time and eye stationary time to obtain the translation characteristic value of the current environment; standardization is a technique in data preprocessing, which realizes the unified measurement between different variables, and the standardized value can be obtained by calculating the mean and standard deviation of the historical data collected in the past period of time.

[0030] Through the multi-sensor fusion technology, the user's head stability, eye gaze state and spatial behavior change are accurately captured, and the real-time dynamic perception of the user's attention and environmental interaction state is realized. The generalized mean aggregation operator is used to weight and fuse the standardized residence time and eye stationary time, which enhances the expression ability of the translation characteristic value and improves the sensitivity and accuracy of the translation demand judgment. Combined with the depth perception distance change, the interaction distance between the user and the translation object is finely monitored, which effectively avoids misjudgment and resource waste, significantly improves the response speed and intelligent level of the translation system, enhances the naturalness and comfort of the user's interactive experience, and is especially suitable for complex and variable actual application scenarios, which helps to improve the market competitiveness and practical value of the smart glasses.

[0031] In step S3, based on the translation characteristic value of the current environment and the distance change of the marked object, a logistic regression model is used to determine whether to translate the marked object.

[0032] First, the translation characteristic value of the current environment and the distance change of the marked object are standardized respectively, and the result of the standardized translation characteristic value of the current environment is denoted as The result of the standardized distance change of the marked object is denoted as .

[0033] The normalized result of the translation feature value of the current environment and the normalized result of the distance change amount of the marked article are taken as input variables of the logistic regression model, and the output is the translation probability value of the marked article, and the mathematical expression form of the logistic regression model is as follows: ; wherein, represents the translation probability value of the marked article, is the normalized result of the translation feature value of the current environment, is the normalized result of the distance change amount of the marked article, is the bias term of the logistic regression model, and are weight parameters of the normalized result of the translation feature value of the current environment and the normalized result of the distance change amount of the marked article respectively, which are obtained by pre-training the logistic regression model based on historical sample data, represents an exponential operation function.

[0034] The translation probability value of the marked article is determined with a preset translation determination threshold value; When the translation probability value of the marked article is greater than or equal to the translation determination threshold value, it is determined to translate the marked article; Otherwise, it is determined not to translate the marked article.

[0035] It needs to be explained that the translation determination threshold value is set by professionals in the field based on the optimization requirements of historical translation behavior data distribution characteristics and translation trigger accuracy indicators, which will not be repeated here.

[0036] The moment when it is determined to translate the marked article is taken as the starting moment of the preset time window, the preset time window is set as the translation time, and the state data of the marked article including the text coverage rate and the article moving speed are collected and detected in real time within the translation time.

[0037] The pixel area of the text area in the image is extracted by the image segmentation algorithm, and the ratio of the pixel area of the text area to the total pixel area of the image is taken as the text coverage rate.

[0038] The article moving speed refers to the position change amount of the marked article in the user's field of view within the translation time, and the linear speed is calculated based on the image center position coordinate difference and the time interval of the marked article within the translation time, and the specific calculation formula is as follows: ;

[0039] wherein, is the article moving speed, and are the image coordinate center points of the marked article at the start and end moments of the translation time, respectively, To translate time.

[0040] It should be noted that the image segmentation algorithm is used to divide the image into several regions with similar attributes for higher-level image understanding, analysis or recognition, in this case, for identifying the pixel area of the character region in the image.

[0041] The use of a logistic regression model for accurate determination of translation requirements enables dynamic identification of user focus targets and scientific decision-making for translation triggering. By standardizing the translation feature values and distance changes, the uniform scale of the input data and the stability of the model are ensured, and the accuracy of translation determination is improved. Real-time collection and analysis of the text coverage and moving speed of the marked object effectively reflect the changing trend of the translation content and the persistence of user attention, further optimizing the translation time window setting and avoiding invalid or redundant translation. This method significantly improves the response efficiency and resource utilization of the intelligent glasses translation system, enhances the system's adaptability to changing information in complex environments, improves the natural smoothness of user experience and the practical value of translation content, and promotes the application and development of intelligent wearable devices in the field of multilingual interaction.

[0042] In step S4, the translation shallowing trend is calculated according to the text coverage and the object moving speed. The greater the text coverage, the longer the text output in the translation bar stays, the slower the translation shallowing trend, and the smaller the value. The greater the object moving speed, the shorter the text output in the translation bar stays, the faster the translation shallowing trend, and the greater the value. The specific calculation expression is as follows: wherein, is the translation shallowing trend, is the text coverage, is the object moving speed, is a small constant to prevent the denominator from being zero.

[0043] Detect the translation language displayed in the translation bar, count the number of characters in the text output in the translation bar, obtain the length of the translation language in the translation bar, and call the preset translation shallowing time benchmark value. The translation shallowing time benchmark value is the basic display time of the text output in the translation bar in the state of a stationary object, which is set by professional personnel based on statistical analysis of a plurality of language display experimental data of stationary objects and translation output readability evaluation results.

[0044] Take the translation language length and the translation shallowing trend as input variables, and use the standardized compression division function to calculate the translation shallowing time correction ratio. The calculation method is as follows: ; wherein, is the translation shallowing time correction ratio, and satisfies , With the increase of the With the increase of the With the increase of the is the translation language length, is the translation shallowing trend, is a tiny constant for preventing the denominator from being zero, and the translation shallowing time correction ratio is used to adjust the time weight of the translation shallowing time under different content complexity and text length.

[0045] Based on the translation shallowing time correction ratio and the translation shallowing time reference value, the two are multiplied to generate the translation shallowing time of the marked item. The translation shallowing time controls the dwell time of the output text of the translation bar, ensuring that users can obtain a more adaptive and reasonable information load translation reading experience under different content density and dynamic conditions.

[0046] By comprehensively considering the text coverage rate and the item moving speed, the translation shallowing trend is scientifically calculated, and the dynamic adaptive adjustment of the text dwell time of the translation bar is realized. This mechanism effectively avoids the problem of user information overload or understanding lag caused by long translation content or fast item movement, and improves the readability of the translation information and the visual comfort of the user. The translation shallowing time correction ratio is calculated by using a standardized compression division function, so that the translation display time can be reasonably adjusted according to the language length and the content, and the adaptability of the system to different scenes and content complexity is enhanced. The overall technical scheme improves the intelligent level and interaction experience of the intelligent glasses translation system, which helps users to obtain more efficient, accurate and cognitive habit-compliant translation services in complex and variable environments.

[0047] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product.

[0048] Those skilled in the art can realize that the modules and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized by hardware or software depends on the specific application of the technical solution and the constraints of the invention. Professional technicians can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0049] In addition, the functional modules in each embodiment of the present application can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0050] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present application, and all of them should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0051] Finally, the above merely provides the preferred embodiments of the present application, and is not used to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A smart glasses based translation method, characterized in that: The method comprises the following steps: Step S1: receiving and recording the native language type input by the user, marking the objects in the current environment by the recognition accessory in the glasses, identifying the information of the marked objects, and determining whether to enter the translation analysis system according to the information identification result and the recorded native language type; Step S2: when entering the translation analysis system, monitoring the residence time of the current environment and the eye ball static time length, fusing the residence time and the eye ball static time length to generate the translation feature value of the current environment, monitoring the distance between the user and the marked objects and counting the distance change amount of the marked objects; Step S3: comprehensively determining whether to translate the marked objects according to the translation feature value of the current environment and the distance change amount of the marked objects, setting the translation time, and detecting the state data of the marked objects within the translation time; Step S4: obtaining the translation shallowing trend according to the state data of the marked objects, detecting the translation language length of the translation column, calling the translation shallowing time reference value, comprehensively calculating the translation shallowing time correction proportion according to the translation language length of the translation column and the translation shallowing trend, and combining the translation shallowing time reference value to generate the translation shallowing time.

2. The translation method based on smart glasses according to claim 1, characterized in that: in step S1, the text content information of the marked objects is identified, the text content information is matched with the language type, and the language type of the marked objects is obtained; if the language type of the marked objects is inconsistent with the native language type input by the user, the translation analysis system is entered, otherwise, the translation analysis system is not entered.

3. The translation method based on smart glasses according to claim 1, characterized in that: in step S2, the time point when the translation analysis system is entered is taken as the residence start time point; a time interval is preset, the head stability angles at the start and end times of the time interval are collected, the difference between the head stability angles at the start and end times of the time interval is taken as the angle change rate, and the ratio of the angle change rate to the time interval is taken as the angle change rate; if the angle change rate exceeds the preset angle change rate threshold, it is determined that the user's residence ends, and the residence stop time point is taken as the residence end time point; the difference between the residence start time point and the residence stop time point is the residence time of the current environment.

4. The translation method based on smart glasses according to claim 3, characterized in that: in step S2, the motion state of the user's eye ball is monitored in real time to determine whether the user's eye ball is in a static staring state; when the movement of the eye ball is monitored, the time point when the movement occurs is recorded; the difference between the residence start time point and the time point when the movement occurs is taken as the eye ball static time length; after the residence time of the current environment and the eye ball static time length are standardized, the standardized residence time of the current environment and the standardized eye ball static time length are obtained respectively; the translation feature value of the current environment is generated by fusing the standardized residence time of the current environment and the standardized eye ball static time length through a generalized mean aggregation operator.

5. The translation method based on smart glasses according to claim 1, characterized in that: In step S2, the distance between the user and the marked object is monitored in real time, and the corresponding marking time point is recorded when the object is marked, and the distance between the user and the marked object at the marking time point is taken as the initial distance; The distance between the user and the marked object at the current time is taken as the current distance; The difference between the initial distance and the current distance is taken as the distance change of the marked object.

6. The translation method based on smart glasses according to claim 5, characterized in that: In step S3, the translation feature value of the current environment and the distance change of the marked object are standardized; The standardized results of the translation feature value of the current environment and the distance change of the marked object are taken as the input variables of the logistic regression model, and the translation probability value of the marked object is output by the logistic regression model.

7. The translation method based on smart glasses according to claim 6, characterized in that: In step S3, when the translation probability value of the marked object is greater than or equal to the preset translation judgment threshold, it is determined that the marked object is translated; Otherwise, it is determined that the marked object is not translated.

8. The translation method based on smart glasses according to claim 7, characterized in that: In step S3, the time when it is determined to translate the marked object is taken as the starting time of the preset time window, and the preset time window is set as the translation time; In the translation time, the pixel area of the text region in the image is extracted, and the ratio of the pixel area of the text region to the total pixel area of the image is taken as the text coverage rate; Based on the image center position coordinate difference of the marked object in the translation time and the translation time, the linear velocity is calculated to obtain the object moving speed.

9. The translation method based on smart glasses according to claim 8, characterized in that: In step S4, the translation shallowization trend is obtained by integrating the object moving speed and the text coverage rate; The number of characters in the translation column output text is counted as the translation language length; The translation language length and the translation shallowization trend are taken as the input variables, and the translation shallowization time correction ratio is calculated by using the standardized compression division function; The translation shallowization time correction ratio is multiplied by the preset translation shallowization time benchmark value to obtain the translation shallowization time of the marked object.

Images